data-transformation-layer
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides documentation and patterns for secure data handling. It specifically mitigates data exposure risks by instructing the agent to use explicit allowlists (manually selecting fields) rather than blocklists when transforming database models to API DTOs.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external data but provides robust architectural mitigations to ensure data integrity.
- Ingestion points: Data is ingested from database records in
mappers/user.tsand from external API responses inhooks/useUser.ts. - Boundary markers: The architecture enforces a strict boundary where data must pass through mapping functions (
toUserResponse,toUserUIModel) before being used by other layers. - Capability inventory: The skill is restricted to data transformation and standard UI fetching; it contains no subprocess calls, dynamic code execution, or file-writing capabilities.
- Sanitization: The skill explicitly demonstrates sanitization by normalizing null values and omitting sensitive fields (like passwords) during the transformation process.
Audit Metadata